Unplanned readmission after surgery for spinal metastases: The influence of data source on a surgical oncology quality indicator - A systematic review and descriptive meta-analysis.

Surgery for spinal metastases represents a complex intervention at the intersection of oncology, neurosurgery, and orthopaedics, performed in patients with advanced systemic disease and limited physiological reserve. Unplanned hospital readmission has emerged as an important quality indicator in surgical oncology, reflecting both perioperative care quality and the trajectory of underlying malignancy. We conducted the first systematic review and descriptive meta-analysis of readmission incidence and risk factors in this population. Five databases were searched (January 2010-March 2026); 14 cohort studies (8132 patients for 30-day analysis; 22,198 for 90-day analysis) met inclusion criteria. Given anticipated heterogeneity between institutional cohorts and national administrative databases, a stratified analytical framework was pre-specified. The exploratory pooled 30-day readmission incidence was 16.0% (95% CI: 12.2-20.3%; 95% prediction interval: 5.3-30.8%; I2 = 93.2%). Stratification revealed homogeneous estimates from single-centre studies (13.9%; I2 = 0%; k = 5) versus higher rates from administrative databases (20.4%; k = 2; p < 0.001), indicating that the data source is a principal driver of the observed heterogeneity. Pooled 90-day incidence was 31.2% (95% CI: 26.9-35.6%; k = 7). Comorbidity burden, prior spinal radiation, and poor functional status were the most consistently identified risk factors, although effect-measure heterogeneity precluded quantitative pooling. Evidence certainty was very low (GRADE). Approximately one in six cancer patients experiences unplanned readmission within 30 days of spinal metastasis surgery, with the wide prediction interval underscoring substantial setting-dependent variability. These findings support the integration of readmission tracking into spinal metastasis surgery quality programmes and prospective validation of risk-stratified perioperative pathways.